Skip to content

File handling

Converted here.
Never uploaded.

There is no conversion server, no storage bucket and no third-party API. The page is static; the parsing happens in your browser, and closing the tab is the deletion step.

Files uploaded

Never

Conversion servers

None

Accounts

None

Ad trackers

None

The path of one file

From drop to tab close

  1. 01

    You pick a file

    The browser opens the file locally. Nothing is posted, uploaded or streamed to this site — there is no conversion endpoint to send it to.

  2. 02

    A worker parses it

    PDF, DOCX and XLSX files go to a dedicated Web Worker so the tab stays responsive. HTML, CSV, JSON and Markdown are read on the main thread by small, local parsers.

  3. 03

    The result stays in the tab

    The Markdown or exported PDF/DOCX is held in page memory for editing and download. It is not written to a server, a database, a queue or a cache the site controls.

  4. 04

    Closing the tab clears it

    Workers are terminated after 60 seconds idle, and the browser releases the page heap when you navigate away. There is no copy left to retrieve.

Facts

The plain list

These are design constraints, not marketing lines. An upload would need an endpoint, a runtime and a retention policy; none of those exist in this project.

Files uploaded to a server
Never
Files sent to third-party APIs
Never
Where conversion runs
This browser tab
Accounts or profiles
None
Analytics or ad trackers
None
Cookies set by the site
None
Transport
HTTPS for the static assets only
Retention window
Zero — tab memory only
Training on your data
Not applicable, nothing is sent

Limits of the guarantee

What this does not promise

Your browser is still software

Extensions with page access, a compromised browser build or a shared device can see what the page sees. That boundary belongs to the device, not to this site.

Memory is finite

A 25 MB binary can expand many times over while it is parsed. The size ceiling, one-at-a-time binary conversion and the mobile warning reduce the risk, but a constrained tab can still run out of memory.

Hosting logs still exist

The static host may keep short-lived request logs for the assets it serves. Those requests do not contain your document, because the document is never sent.

Questions

File handling FAQ

So my file is never uploaded?

Correct. There is no upload endpoint in this build. The file is read with the browser File API, parsed in a Web Worker or on the page, and released when you close the tab.

Do you log IP addresses?

The application has no server, no cookies and no analytics. The static host (Cloudflare Pages) may keep ordinary edge logs for operations and abuse prevention, but those requests only fetch HTML, JavaScript, CSS and fonts — never document contents.

Can a file be recovered after conversion?

No. The file never leaves your device, so there is no server archive to recover. The output in your browser is the only copy the converter produced.

What about sensitive documents?

Local processing removes the upload risk, but it does not make a hostile browser safe. Extensions with page access can read what the page can read. For material that must never touch a browser extension, use a local command-line tool such as pandoc or a desktop app.

Is the site open to audit?

The conversion code is ordinary TypeScript in the repository, including the worker entry points and the pure readers. There is no hidden upload step and no server route: you can inspect every parser that runs.

What are the remaining risks?

Large files can exhaust a browser tab’s memory, a compromised extension can read the page, and a malicious document could target an upstream parser bug. The 25 MB ceiling, one-at-a-time binary conversion and the mobile warning reduce the first risk. SECURITY.md in the repository records the full matrix and the residual risks.

Convert without leaving a copy behind

Drop a file, take the result, close the tab. The file never leaves your device.